What Food Crisis? Measuring Global Hunger

Timothy A Wise

Last week, Oxfam launched its new international campaign, GROW, to fight food insecurity. The advocacy organization’s campaign materials cite many of the statistics with which the post-food-crisis world has become familiar. Most common is the estimate that more than one billion people in the world are now hungry as a result of the combined impacts of rising food prices and the global economic recession. The estimate comes from the UN’s Food and Agriculture Organization (FAO), and few have questioned the validity of the numbers.

Now two studies suggest the estimate may be inflated. In the May/June special issue of Foreign Policy magazine on food, Abhijit Bannerjee and Esther Duflo, from their perches at MIT’s Poverty Lab Project, have an article with the provocative subtitle, “but what if the experts are wrong?” Meanwhile, IFPRI’s Derek Headey, in a VoxEU post, examines the prevailing FAO/World Bank methodologies for estimating global hunger and suggests that these institutions are overestimating hunger, mainly because they discount the positive impacts of economic growth in some of the world’s most populous countries.

On closer inspection, Bannerjee and Duflo deepen our understanding of the nature of hunger in developing countries, but they offer little here to call into question the billion-hungry estimate. Headey, on the other hand, is onto something, but it’s worth going deeper still to understand the relationship between poverty, high food and agricultural prices, economic growth, and government policy.

Bannerjee and Duflo, using their highly empirical Poverty Lab methodologies, really aren’t trying to answer the question of how many hungry people there are on the planet. Rather, they make us look more closely at the nature of hunger and poverty, pointing out that simple economic assumptions about the poor’s food-buying and food-consuming habits are fraught with misconceptions. For example, they assert from their field experience that the poor choose foods not only or even primarily based on cost and nutritional value but based on how good they taste. This is a worthwhile read, but this in no way answers FP’s provocative title questioning the prevalence of hunger.

Headey is more on-target with his critique of the global numbers. In his VoxEU summary, a post on Dani Rodrik’s blog, and in a longer paper, he compares the common FAO/WB method for estimating hunger using simulation analysis based on caloric intake with his own culling of international Gallup polling data on self-reported hunger and food insecurity. His results are striking, suggesting that from 2005/6 to 2007/8, when agricultural prices skyrocketed, the number of hungry in the 70 countries for which there was data declined by 408 million.

How could this be? I’m not going to summarize his full argument and data here; he does a nice job of that on his VoxEU post. But he claims that the Gallup data show that in the largest developing countries with the largest number of poor and hungry, notably China and India, economic growth more than made up for any negative impacts from rising food prices.

Headey calls for a re-examination of the methodologies for estimating global hunger, and his data certainly justify that call. That said, China has been the dragon in the room for a lot of statistical anomalies in recent years, with its fast growth, declining poverty, and large population swamping generalizations about progress in “the developing world.” Headey notes that China accounts for two-thirds of the decline in self-reported food insecurity, and a closer look at his data suggests that a large portion of developing countries experienced an increase in hunger in that same period. So for those who might take this data as a reason to do less about global food insecurity: not so fast!

There are other important unanswered questions in Headey’s study:

  1. The period he studied did not fully account for the impact of the economic downturn on the global economy. As such, he was measuring the impact of food price inflation, not its toxic combination with slow or negative economic growth. That’s what the poor have faced; to deal with global hunger, that’s what we need to understand.
  2. There were important policy reasons that high agricultural prices did not have as big an impact in some parts of the world, notably China and India. Both are relatively self-sufficient in their most basic foods, and both maintained buffer stocks with which they could offset rising international prices. So one would expect to see lower impacts of rising food prices because, well, food prices didn’t rise as much. How does his analysis account for this?
  3. I’ve argued in an earlier blog post that high agricultural prices can be good for development and poverty/hunger in countries where a large share of the poor work in agriculture. Countries such as China and India. So to what extent is the decline in self-reported food insecurity reflecting the positive economic-stimulus effects of higher prices– in farmer incomes and returns to unskilled labor in and out of agriculture?

Headey starts an important discussion. Hopefully it takes place without undercutting concern for global food insecurity and the worthy campaigns to address it, such as Oxfam’s.

4 Responses to “What Food Crisis? Measuring Global Hunger”

  1. Hi Tim,

    Thanks for covering this. It’s important. And I think you have touched on the key issues the paper raised for me. One, I learned about how the numbers are derived. I had no idea – and the method is clearly very imperfect. Then, I was less persuaded by the conclusions, because India and China are the two reasons global statistics are often not much use. As you say, averaging them in skews the numbers for all the other 170+ non-OECD countries.

    Also, we are going to have to disentangle the 2007/08 crisis from the current situation, which is clearly not unrelated to the what happened three years ago, but which is not a replay either. My understanding is that the transmission of high global prices to domestic economies was very uneven, and in some cases slow. However, transmission was complete. And when prices came down globally, they did not come down in developing countries anything like as fast or as far. This time, with the renewed price rises starting in 2010, the transmission has been faster. Understanding these dynamics will be a country by country and region by region task. It’s important to do the analysis.

    Clearly, it is a complicated and dynamic situation. As you say, relative prosperity in India and China should not confuse policy-makers as to the gravity of the situation faced by a great number of people living in poverty. We want higher prices, but

    I appreciated Headey raising the question about the benefits of higher commodity prices for countries that are largely dependent on commodity revenues. Presumably, despite the gross inequities of market power, higher cotton, coffee and also cereal prices should be good for rural areas. They may hurt consumers in the short-run, but better prices at the farmgate usually improve wages for labour and returns for producers and therefore demand for services, including schools. This is a dynamic effect that we want. The challenge is in how to make it more stable: high prices in the context of the volatility world markets are now experiencing is a very uncomfortable place to be for those with scarce resources.

    Time for another post on how to calm the volatile waters? Unfortunately, I don’t think we need hold our breath to see what the summit of G20 Agriculture Ministers will come up with when they meet on 23 June in Paris. The answer is: far too little.

  2. Jayati Ghosh says:

    Comment on Tim Wise blog
    Many thanks, Tim, for this very important critique of a potentially misleading position that is suddenly finding more takers in the international development community.
    It is worth looking carefully at the Gallup Poll methodology before we decide to jump to conclusions. The Gallup report on its food security survey notes that it is based on telephone and face-to-face interviews conducted throughout 2005, 2006, 2007, and 2008, with randomly selected sample sizes (typically around 1,000 residents) in 134 countries – so a total of less than 140,000 people across the world. The distribution of the samples across urban and rural locations or by income category is not clear at all, nor is the proportion that was contacted by telephone. This is not exactly a solid basis on which to draw major conclusions on the extent of global hunger. The Gallup Poll people themselves do not seem to think they can make intertemporal comparisons based on these data: their own conclusion is that “even before the crisis, affording food was a challenge for many”.
    Basing a major conclusion on this rather weak “self-perception” data, as Headey does, is really not justifiable. Instead, consider the main mechanism through which higher global food prices affect people: domestic food prices. Here, as Sophia points out, the disturbing news is that the transmission of increases in food prices has generally been rapid (and is getting faster and more complete) while the downward movements have not been transmitted so much.
    The FAO’s Global Food Price Monitor for June 2011 suggests that domestic prices are affected by international prices, but the extent also depends upon the degree of domestic self-sufficiency and the behavior of local harvests. Recent good harvests for wheat in Asia have helped to restrain domestic wheat prices in the region even though global trade prices of wheat have increased dramatically. By contrast, Sub-Saharan maize importing countries have experience sharp increases in domestic maize prices.
    None of this should come as a big surprise. What is more surprising is that as the standard indicators of nutrition deteriorate as food prices increase, the attempt is made to move the goal posts, by claiming that these indicators do not matter as much as people’s perceptions (as reported in relatively small surveys that are not intertemporally comparable) or people’s choices about how to spend their incomes. If nutrition and hunger are genuinely seen to be problems, they need to be addressed directly without such obfuscation. Tim’s blog is an important step in that direction.

  3. Derek Headey says:

    Dear Timothy & other bloggers,
    First, many thanks for your interest in my paper. I only have a couple of corrections. First, Although it is strictly true that the raw Gallup data suggest that over 400 million people were thrown out of self-reported food insecurity, I don’t emphasize that number if my paper because there are major doubts about the validity of the Chinese data for 2006. That’s why I carry out a number of sensitivity analyses based on alternative estimates for trends in China. The gist of those alternatives estimates is that food security probably went down at a global level, though it is hard to say by how much.
    Second, I don’t have much to say about the financial crisis impacts, although the financial crisis coincided with falling food prices which presumably helped many poor people cope with any adverse income effects (though the latter are not well quantified either).
    Third, the Gallup methodology does indeed deserve much more attention (although they don’t do phone interviews in developing countries, only richer countries). Their samples are intended to be nationally representative, but they do have substantial margins of error, and some peculiar outliers. The poll would benefit from more tests (such as test-retests), more validation with other sources, and more disaggregation (rural-urban, farm-nonfarm). But the real point of my paper is that their numbers cast justifiable doubt on the FAO and World Bank numbers, and I think a number of top academics will agree. Readers might also be interested in a recent piece by Mason and Jayne of Michigan State. They looked at urban wage and food price data to see if relatively poor urban people in East Africa were hard hit by higher food prices. The answer is that they were, but that real wage growth over the course of anumber of years meant that these folks were still better off at the height of the crisis than they were several years ago. Those results are quite consistent with my own.
    Moving forward, what we need to do now is to think hard about how we can improve on the measurement of food insecurity. The crisis revealed some major weaknesses in existing efforts, and it’s not yet clear where we should be heading in this regard.

  4. Timothy A. Wise says:

    Thanks to Derek for the helpful and detailed response. Note that Will Martin and Hassam Zaman pose questions about the comparability of the Gallup methodology with the simulation results, in a guest post on Dani Rodrik’s blog. They conclude, in part:
    “Even with these methodological differences, we still find the GWP result of a decline of 400 million in food-insecure people between 2005-6 and 2007-8 surprising. This is a large decline when we know that global food prices spiked in 2007-8; that the poorest households spend 60-70% of their income on food; and that energy prices were also rising….”
    See the post at: